Image Classification on ImageNet (val) (Top-1 and Top-5 Error)
21Top-1 ErrorHRNet-W96-C
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HRNet-W96-C#Params.=57.5M, GFLOPS=10.2, Residual branch type=bottleneck2019.08 | 21 | 5.6 | |
| HRNet-W40-C#Params.=57.6M, GFLOPS=11.8, Residual branch type=two 3 x 3 convolutions2019.08 | 21.1 | 5.6 | |
| ResNet-152#Params.=60.2M, GFLOPS=10.7, Residual branch type=bottleneck2019.08 | 21.2 | 5.7 | |
| HRNet-W76-C#Params.=40.8M, GFLOPS=7.3, Residual branch type=bottleneck2019.08 | 21.5 | 5.8 | |
| ResNet-101#Params.=44.6M, GFLOPS=7.3, Residual branch type=bottleneck2019.08 | 21.6 | 5.8 | |
| HRNet-W30-C#Params.=37.7M, GFLOPS=7.55, Residual branch type=two 3 x 3 convolutions2019.08 | 21.9 | 5.9 | |
| ResNet-106#Params.=64.9M, GFLOPS=11.1, Residual branch type=two 3 x 3 convolutions2019.08 | 22.7 | 6.4 | |
| HRNet-W44-C#Params.=21.9M, GFLOPS=3.9, Residual branch type=bottleneck2019.08 | 23 | 6.5 | |
| HRNet-W18-C#Params.=21.3M, GFLOPS=3.99, Residual branch type=two 3 x 3 convolutions2019.08 | 23.1 | 6.5 | |
| ResNet-72#Params.=48.4M, GFLOPS=7.46, Residual branch type=two 3 x 3 convolutions2019.08 | 23.3 | 6.7 | |
| ResNet-50#Params.=25.6M, GFLOPS=3.82, Residual branch type=bottleneck2019.08 | 23.3 | 6.6 | |
| ResNet-38#Params.=28.3M, GFLOPS=3.8, Residual branch type=two 3 x 3 convolutions2019.08 | 24.6 | 7.4 |